We are trying to solve this problem for months, but we have tried all the solutions from the web and the problems remains, also in models that where correctly working.
Our company is deploying W7 64 bit, on a lot of different models, using Landesk, using many vbs scripts made by us. The gold image is made on a virtual machine, then syspreped and captured. This syspreped wim is deployed to the client by a WinPE 5.0 x86 (W8.1 32 bit), using imagex command, the same problems happens using dism commands. First of all, we partition the disk in MBR, with one partition of 300 mb for booting, and another one for the system. Then the syspreped wim is applied, to both partitions (we have also tried to create de boot partition instead of appliying it). At last, we install the drivers to the offline image using dism. So, although we use Landesk, the problem happens doing all the process manually with WinPE, and I think it would happen with any deployment app.
It happens on new and older machines, at the beginning we think that it could be because of the SSD disks, but it also occurs on non SSD Disks…We use use BIOS mode, instead of UEFI, but we have tried with this one, and is the same.
Finished this process, the machine is restarted, but when Windows logo is charging it freezes, and gives a BSOD with 0x7E error. We have seen that this only happens with a fully updated image, not if we use a clean W7 installation captured, but when we update this virtual machine, and capture it, always happens. The only way we can avoid this, is not installing the drivers, or remove some drivers after a lot of test-error. It always freezes when loading disk.sys (seen at safe mode) but, sometimes is due to the audio driver, another one the usb 3, another one the hotkeys… it has no sense, and we have not found any easy way to know it, because no log or dump is created.
We have installed the NVMe patches from Microsoft on the image, installed the intel and Samsung NVMe disk on the offline image, tried not installing some of the Microsoft updates on the image, making the image from the beginning again, but nothing has been successful. And to make it more strange, sometimes it happens on the same machine at the first deployment, but if you try it second, it works…
Can anybody see the sense of this behavior? Any idea of how to solve this?
Thanks in advance.